Value-Function-Based Transfer for Reinforcement Learning Using Structure Mapping

By Yaxin Liu, Peter Stone. 2006

92 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Reinforcement Learning

Citations in the corpus

Learning Parameterized Skills (2012, methodology): “Liu and Stone (2006) propose a method for transferring a value function between pairs of tasks but require prior knowledge of the task dynamics in the form of a Dynamic Bayes Network.”
Multi-agent reinforcement learning as a rehearsal for decentralized planning (2016, positive): “The key challenge is to map the value functions between the two different tasks [16], which we accomplish for RLaR by mapping ðs; t; a !Þ to ðs;hi; a !Þ such that t 1⁄4 jhi j=2þ1, i.”
Grounding Language for Transfer in Deep Reinforcement Learning (2017, unknown): “Most previous approaches have either explored skill transfer (Konidaris & Barto, 2007; Konidaris, Scheidwasser, & Barto, 2012) or value function/policy transfer (Liu & Stone, 2006; Taylor, Stone, & Liu, 2007; Taylor & Stone, 2007).”
Deep Transfer in Reinforcement Learning by Language Grounding (2017, background): “Most previous approaches have either explored skill transfer (Konidaris and Barto, 2007) or value function transfer (Liu and Stone, 2006).”
Unsupervised Cross-Domain Transfer in Policy Gradient Reinforcement Learning via Manifold Alignment (2015, background): “…A PRELIMINARY VERSION OF THIS PAPER THE FINAL VERSION WILL APPEAR IN THE AAAI DIGITAL LIBRARY tic knowledge about state features between two tasks (Liu & Stone 2006), background knowledge about the range or type of state variables (Taylor et al. 2007b), or transition models for each possible…”
Cross-domain transfer for reinforcement learning (2007, background): “If a human can de.ne both tasks can be de.ned in terms of qualitative dynamic Bayes networks, Liu and Stone (2006) showed that a graph matching method can automatically .nd task similarities.”
Autonomous transfer for reinforcement learning (2008, background): “[20] M. E. Taylor, P. Stone, and Y. Liu.”
Reinforcement Learning Transfer Using a Sparse Coded Inter-task Mapping (2011, background): “For example, semantic knowledge about state features between two tasks may be used [8, 5], background knowledge about the range or type of state variables can be used [13, 17], or transition models for each possible mapping could be generated and tested [14].”
An Introduction to Intertask Transfer for Reinforcement Learning (2011, background): “For instance, Liu and Stone (2006) assume that a type of qualitative model is available for learning a mapping, but such a model is insufficient for planning a policy.”
Transfer Learning via Inter-Task Mappings for Temporal Difference Learning (2007, background): “Liu and Stone (2006) define qualitative dynamic Bayes networks which summarize the effect of actions on different state variables.”
Exploiting Agent and Type Independence in Collaborative Graphical Bayesian Games (2011, methodology): “n spaces, as done in transfer learning (Taylor et al., 2007). Furthermore, it may be possible to automatically identify suitable source tasks and mappings between tasks, e.g., using qualitative DBNs (Liu and Stone, 2006). 7 Conclusions In this article, we considered the interaction of several agents”
Automatically Mapped Transfer between Reinforcement Learning Tasks via Three-Way Restricted Boltzmann Machines (2013, background): “For example, semantic knowledge about state features between two tasks may be used [8,10], background knowledge about the range or type of state variables can be used [14,19], or transition models for each possible mapping could be generated and tested [16].”
Transfer Learning for Reinforcement Learning Domains: A Survey (2009, background): “For instance, qualitative dynamic Bayes networks (QDBNs) (Liu and Stone, 2006), summarize the effects of actions on state variables but are not precise (for instance, they could not be used as a generative model for planning).”
Transfer in variable-reward hierarchical reinforcement learning (2008, methodology): “This work has been extended to automate the determination of the similarity mappings (Liu and Stone 2006).”